A quasi-locally most powerful test for correlation in the conditional variance of positive data

Brendan McCabe, Gael Martin, Keith Freeland

Research output: Contribution to journalArticleResearchpeer-review

Abstract

A test is derived for short-memory correlation in the conditional variance of strictly positive, skewed data. The test is quasi-locally most powerful (QLMP) under the assumption of conditionally gamma data. Analytical asymptotic relative efficiency calculations show that an alternative test, based on the first-order autocorrelation coefficient of the squared data, has negligible relative power to detect correlation in the conditional variance. Finite-sample simulation results confirm the poor performance of the squares-based test for fixed alternatives, as well as demonstrating the poor performance of the test based on the first-order autocorrelation coefficient of the raw (levels) data. The robustness of the QLMP test, both to misspecification of the conditional distribution and to misspecification of the dynamics, is also demonstrated using simulation. The test is illustrated using financial trade durations data.
Original languageEnglish
Pages (from-to)43 - 62
Number of pages20
JournalAustralian and New Zealand Journal of Statistics
Volume53
Issue number1
DOIs
Publication statusPublished - 2011

Cite this

@article{4625fa3fda4349758f282559caaeb38f,
title = "A quasi-locally most powerful test for correlation in the conditional variance of positive data",
abstract = "A test is derived for short-memory correlation in the conditional variance of strictly positive, skewed data. The test is quasi-locally most powerful (QLMP) under the assumption of conditionally gamma data. Analytical asymptotic relative efficiency calculations show that an alternative test, based on the first-order autocorrelation coefficient of the squared data, has negligible relative power to detect correlation in the conditional variance. Finite-sample simulation results confirm the poor performance of the squares-based test for fixed alternatives, as well as demonstrating the poor performance of the test based on the first-order autocorrelation coefficient of the raw (levels) data. The robustness of the QLMP test, both to misspecification of the conditional distribution and to misspecification of the dynamics, is also demonstrated using simulation. The test is illustrated using financial trade durations data.",
author = "Brendan McCabe and Gael Martin and Keith Freeland",
year = "2011",
doi = "10.1111/j.1467-842X.2010.00596.x",
language = "English",
volume = "53",
pages = "43 -- 62",
journal = "Australian and New Zealand Journal of Statistics",
issn = "1369-1473",
publisher = "Wiley-Blackwell",
number = "1",

}

A quasi-locally most powerful test for correlation in the conditional variance of positive data. / McCabe, Brendan; Martin, Gael; Freeland, Keith.

In: Australian and New Zealand Journal of Statistics, Vol. 53, No. 1, 2011, p. 43 - 62.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - A quasi-locally most powerful test for correlation in the conditional variance of positive data

AU - McCabe, Brendan

AU - Martin, Gael

AU - Freeland, Keith

PY - 2011

Y1 - 2011

N2 - A test is derived for short-memory correlation in the conditional variance of strictly positive, skewed data. The test is quasi-locally most powerful (QLMP) under the assumption of conditionally gamma data. Analytical asymptotic relative efficiency calculations show that an alternative test, based on the first-order autocorrelation coefficient of the squared data, has negligible relative power to detect correlation in the conditional variance. Finite-sample simulation results confirm the poor performance of the squares-based test for fixed alternatives, as well as demonstrating the poor performance of the test based on the first-order autocorrelation coefficient of the raw (levels) data. The robustness of the QLMP test, both to misspecification of the conditional distribution and to misspecification of the dynamics, is also demonstrated using simulation. The test is illustrated using financial trade durations data.

AB - A test is derived for short-memory correlation in the conditional variance of strictly positive, skewed data. The test is quasi-locally most powerful (QLMP) under the assumption of conditionally gamma data. Analytical asymptotic relative efficiency calculations show that an alternative test, based on the first-order autocorrelation coefficient of the squared data, has negligible relative power to detect correlation in the conditional variance. Finite-sample simulation results confirm the poor performance of the squares-based test for fixed alternatives, as well as demonstrating the poor performance of the test based on the first-order autocorrelation coefficient of the raw (levels) data. The robustness of the QLMP test, both to misspecification of the conditional distribution and to misspecification of the dynamics, is also demonstrated using simulation. The test is illustrated using financial trade durations data.

U2 - 10.1111/j.1467-842X.2010.00596.x

DO - 10.1111/j.1467-842X.2010.00596.x

M3 - Article

VL - 53

SP - 43

EP - 62

JO - Australian and New Zealand Journal of Statistics

JF - Australian and New Zealand Journal of Statistics

SN - 1369-1473

IS - 1

ER -